Increasing lead base and value proposition of healthcare Scribe
Turi Saúde is an early-stage healthcare startup on a mission to give physicians back their time — by taking the paperwork off their plate. As Founding Designer, I led the end-to-end design, research, and part of the build of all the digital products. Collaborating closely with the Founder and Engineers.

PROBLEM
Interviewed Radiologist
“During the appointment, patients bring lab results printed. So I take pictures and write things down later, otherwise I lose too much time.”
Challenge
The current AI-Scribe product partially addresses the clinical documentation workload.
In a pre-seed startup environment, prioritization is key. Our goal was to validate new features fast, gain users trust within a complex workflow, while optimizing resources.
RATIONALE
Turi's Experimentation Platform
We decided to launch a vibe-coded platform to validate features faster without engineering dependency, given capacity constraints.
In a few months, I designed and launched 4 features based on user interviews, surveys, and feedback, increasing our lead base by 50%.
testing
New feature validated: Lab results summarizer
The feature synthesizes multi-page files into a summary, ready to paste into the EHR.
User interviews revealed:
✅ Strong perceived value by users
✅ Reduced patient anxiety during consultations
⚠️ Improvements needed with OCR model and output experience
Data showed:
7,300+ lab results processed by physicians
Hundreds of registered physicians

TESTING
Potential for market differentiation: LME form auto-fill
The LME form is a mandatory document for approving high-cost medications, with strict completion requirements. Filling errors rank as the third leading cause of deferrals, delaying treatment for millions of patients.
User interviews revealed:
✅ Strong interest by B2B partners
✅ Positive social impact
✅ Not offered by competitors
⚠️ Opportunity to increase auto-fill from 21% to 100%.
Data showed:
820+ forms created by users

SOLUTION
Integrating lab results summarizer with the scribe
With strong perceived value by users and data showing user engagement, I gained stakeholders' buy-in to integrate the validate feature into the scribe.
Data showed:
43,8% click through rate
53% feature adoption rate among Scribe users

DESIGN
Striving for discoverability and mobile responsive
Giving the Scribe existing workflow and ~25% of mobile users:
Replicate upload component from Turi Lab wasn't an option
Analyzing user sessions, I decided to decrecate the text tool bar and reclaim space
Two critical design decisions:
Tag AI output as "context data" → ensures chronological documentation for diagnosis
Enable processing at any point in visit flow → respects varied physician workflows


